I'd need to look into what effect this might have on accessibility, but my gut says "very little".

Many who notice won't bother commenting, because most who notice know how pointless that is (counterproductive in fact: a comment is an interaction, any interaction is a positive for the "content"). Those that do notice and comment are either drowned out by

• those too numbed on the brain to care, let alone notice, who lap it up, and praise it

• bots (either those being used to interact with the clip to drive it's interactions counters, or more general spam bots)

or if there is anyone/anybot monitoring the negative comments are removed.

A written offer on its own would not normally be directly enforceable in many (most?) jurisdictions, for the same sort of reason that retailers can't be held to incorrectly published prices (in the UK at least, a displayed price is an “invitation to tender”, not a contract or other promise) except where other laws/regulations (anti bait&switch rules for instance), or the desire to avoid fighting in the court of public opinion, come into effect.

But in this instance, the written offer and the response to that offer are part of the wider licence that has been agreed to.

> If distribution of object code is made by offering access to copy from a designated place, then offering equivalent access to copy the source code from the same place satisfies the requirement to distribute the source code, even though third parties are not compelled to copy the source along with the object code.

Similar clauses in Sec 6.

[1] https://www.gnu.org/licenses/old-licenses/lgpl-2.1.html

> c) Accompany the work with a written offer, valid for at least three years, to give the same user the materials specified in Subsection 6a, above, for a charge no more than the cost of performing this distribution.

Or, instead of theorycrafting reasons why it shouldn't work, you could "just" sue them and see if the judge agrees.

The hell? Over here, the price tags are a sort of public contract, to which the seller pre-commits. The seller forgot to change the tags? That's not the buyer's problem.

The other solid bait & switch is advertising a product that they don't have any of to sell, in the hope that you'll come in and buy something more expensive (or lower value) instead.

The same as automated apologies.

Not from an “AI”, but I spent over an hour⁰ waiting for a delayed train¹, then the journey, on Tuesday, being regaled every few minutes with an automated “we apologise for your journey taking longer than expected” which is far more irritating than no apology at all.

--------

[0] I lie a little here - living near the station and having access to live arrival estimations online meant I could leave the house late and only be waited on the platform ~20 minutes, but people for whom this train was a connecting leg of a longer journey didn't have that luxury.

[1] which was actually an earlier train, the slot in the timetable for the one I was booked on was simply cancelled, so some were waiting over two hours

That and I'm trying to re-learn Spanish to a useful degree¹ - better attentiveness to the task would help with that too.

--------

[1] I did pretty well at GCSE level three decades ago, then never used it again. I fancy spending time over there in future and don't want to be that guy who turns up and expect everyone to speak some English.

The difference with the 386, I think, is that AFAIK the second-sourced 8086 and 286 CPUs from non-Intel manufacturers still made use of licensed Intel designs. The 386 (and later) had to be reverse engineered again and AMD designed their own implementation. That also meant AMD was a bit late to the game (the Am386 came out in 1991 while the 80386 had already been released in 1985) but, on the other hand, they were able to achieve better performance.

>AMD said Friday that its “independently derived” 486 microprocessor borrowed some microcode from Intel’s earlier 386 chip.

Borrowed hehe. Ended up in a 1995 settlement where AMD fully admitted copying and agreed to pay $58mil penalty in exchange for official license to 386 & 486 microcodes and infamous patent 338(mmu). Intel really wanted a legal win confirming validity of their patent 338 to threaten other competitors. 338 is what prevented sale of UMC Green 486 in USA. Cyrix bypassed the issue by manufacturing at SGS and TI who had full Intel license https://law.justia.com/cases/federal/district-courts/FSupp/8...

>were able to achieve better performance

Every single Am386 instruction executes at same cycle count as Intel counterpart, difference is only official ability to work at 40MHz.

It is, yes. I meant to mention that detail!

> The 386 (and later) had to be reverse engineered … That also meant AMD was a bit late to the game

There were also legal matters that delayed the release of their chips. Intel tried to claim breach of copyright with the 80386 name¹ and so forth, to try stymie the competition.

> they were able to achieve better performance.

A lot of that came from clocking them faster. I had an SX running at 40Hz. IIRC they were lower power for the same clock then Intel parts, able to run at 3.3V, which made them popular in laptops of the time. That, and they were cheaper! Intel came out with a 3.3V model that had better support for cache to compete with this.

--------

[1] This failed, which is part of why the i386 (and later i486 and number-free names like Pentium) branding started (though only in part - starting to market direct to consumers rather than just EOMs was a significant factor in that too).

> It is RECOMMENDED that all senders and recipients support, at a minimum, URIs with lengths of 8000 octets in protocol elements.

It is always worth remembering that, unless you have already ensured that the content has been rendered into a URI-safe subset of ASCII, a character and an octet are not the same thing.